Indirect Reciprocity Game Simulation on Infinite-Horizon Indirect Reciprocity Game
100Cooperation RatioALIGN (DeepSeek-V3.1 Reasoner)
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ALIGN (DeepSeek-V3.1 Reasoner)Agent Type=Reasoning Models, Backbone=DeepSeek-V3.1 Reasoner2026.02 | 100 | 4 | 4 | 15.76 | 0 | |
| ALIGN (GPT-4o Mini)Agent Type=Chat Models, Backbone=GPT-4o Mini2026.02 | 95 | 3.6 | 3.8 | 14.97 | 0.03 | |
| ALIGN (o4-mini)Agent Type=Reasoning Models, Backbone=o4-mini2026.02 | 95 | 3.6 | 3.8 | 14.97 | 0.04 | |
| ALIGN (Kimi-K2-Instruct)Agent Type=Reasoning Models, Backbone=Kimi-K2-Instruct2026.02 | 70 | 1.6 | 2.8 | 11.04 | 0.18 | |
| ALIGN (DeepSeek-V3.1 Chat)Agent Type=Chat Models, Backbone=DeepSeek-V3.1 Chat2026.02 | 68 | 1.4 | 2.7 | 10.63 | 0.2 | |
| ALIGN (Qwen3-235B-Instruct)Agent Type=Reasoning Models, Backbone=Qwen3-235B-Instruct2026.02 | 49 | -0.1 | 1.95 | 7.66 | 0.19 | |
| ALIGN (Gemini 2.5 Flash-Lite)Agent Type=Chat Models, Backbone=Gemini 2.5 Flash-Lite2026.02 | 23 | -2.2 | 0.9 | 3.55 | 0.35 | |
| ALIGN (LLaMA 4 Maverick)Agent Type=Chat Models, Backbone=LLaMA 4 Maverick2026.02 | 17 | -2.6 | 0.7 | 2.75 | 0.87 |